With an increasing requirement for transmission speed and spectral efficiency, communication system may satisfy the requirement for transmission speed and spectral efficiency in a higher order quadrature amplitude modulation (QAM) mode. An important subject faced by industry is that a method for effectively mapping bit information to be transmitted to high order modulation constellation points, so as to ensure spectrum efficiency and satisfy basic requirement of the system for bit error rate (or frame error rate).
In academia, a solution of multilayer coding modulation is provided for this subject. In the solution, transmitted bit streams are demultiplexed and divided into parallel sub information streams; a channel encoding is performed on each of sub information streams; and uniform symbol mapping is perform on bit data in each coded channel to generate a symbol data stream. Based on the solution, in view of realizability and an improvement of frequency efficiency, lots of engineering feasible solutions are provided in the industry. For example, after transmitted bit data is demultiplexed, the channel encoding is performed on only one of parallel channels of bit data, which will not be performed on the other channel of bit data.
In conventional technology, when encoded bit data and uncoded bit data are allocated, information such as a mapping order of high order modulation constellation points, a code rate of channel encoding and a predefined mapping manner, etc., is comprehensively considered to divide the two channels of bit data into important bit data and unimportant bit data. The important bit data is configured to divide constellation points into disjoint subsets of constellation points, called coset, and thus the channel encoding is not performed on the important bit data. However, the channel encoding is performed on the unimportant bit data, and high order modulation constellation points corresponding to the encoded bit data are selected from the coset in the predefined mapping manner.
In foregoing conventional technology, in one aspect, amount of important bits and amount of unimportant bits may be allocated according to practical application scenarios to implement different frequency efficiency. In another aspect, combined with characteristics of a constellation diagram, although the complex channel encoding may not be performed, important bits may reach the same code error rate/frame error rate as those coded with channel encoding, thus reducing implementation complexity. Therefore, in conventional technology, a compromise design is implemented to some extent in view of frequency efficiency and implementation complexity. However, when a high order modulation is demodulated in an existing mapping mode, important bits is only decided based on least Euclidean distance criteria, thereby causing poor decision accuracy for symbol data.